WRITERS’ VOICE AND ENGAGEMENT STRATEGIES IN STUDENTS’ ANALYTICAL EXPOSITION TEXTS

One criterion of a good analytical exposition lies in the writers’ ability to present a clear position and show their voice in their writing. Such quality can be achieved through skillful use of Engagement resources which allow writers to state their voice by aligning or disaligning it with those of others in presenting their case. Involving nine analytical exposition texts written by university students of different proficiency levels, this study explores how the writers’ voice (Hyland, 2008) is constructed through engagement resources. Drawing on Martin and White’s (2005) Engagement system, the study reveals that, while all of the students are capable of presenting a clear position, students of different proficiency levels indicate different engagement strategies in their writing. Students who are more proficient in English are able to more successfully exploit the resources necessary for constructing a well-argued text and show a stronger sense of authorship. This study is expected to give insights into the use of engagement resources in developing the writer’s voice in texts written by EFL writer learners in the Indonesian context.

Academic writing is notably one of the most important skills students need to develop at the tertiary level of education.A good academic writing requires the writer to present a clear position and show engagement with a range of ideas to support it.Consequently, it is of high importance that writers are able to express their voice in their writing.'Voice' here refers to how writers communicate their views and engage their readers with the texts (Hyland, 2008;White, 2006).These texts, therefore, are not merely a mosaic of facts, experts' opinions or conclusions from other texts but, more importantly, they also reflect a clear positioning and identity of the writers (Hyland, 2002).In the field of Systemic Functional Linguistics (SFL), one framework that can be used to analyze writers' voice is 'Appraisal' (Martin & White, 2005).Developed in the 1990s as an extension of the interpersonal meaning of language in SFL, Appraisal is a system that seeks to explain how language is used to evaluate attitude, take stances, create authorial identity, and construct interpersonal positionings and relationships (Hood, 2010;Martin & White, 2005;Wei, Wherrity, & Zhang, 2015).There are three main resources within appraisal: engagement, attitude, and graduation.Due to the study's limited scope, however, only engagement resources are drawn on as analytical device here.Needless to say, the present study is expected to contribute to raising teachers and students' awareness of the importance of voice in writing, as well as to enhancing their writing skills.

Engagement and the writer's voice
Drawing on Bakhtin's and Voloshinov's notions of dialogism and heteroglossia, which highlight the idea that all forms of verbal communication echo voices that have been uttered elsewhere, while simultaneously anticipate possible responses from the audience, Martin and White (2005) developed a system called engagement which extends the interpersonal meaning in SFL.This system covers all "locutions which provide the means for the authorial voice with respect to, and hence to 'engage' with, the other voices or alternative positions construed as being in play in the current communicative context" (p.94).Engagement in written texts thus relates to how writers position themselves with respect to other voices.However, a text, as Martin and White (2005) point out, may also feature bare assertions that overtly present no alternative positions, making the particular communicative event single voiced.The idea of this single-voicedness is referred to as monoglossic, which is opposed to heteroglossic or multivoicedness.Examples of monoglossic and heteroglossic locutions are presented in Table 1 (Adopted from Martin & White, 2005, p. 100).
In overtly dialogistic locutions, Martin and White (2005) distinguish two broad categories based on how other voices and alternative positions are doi: 10.17509/ijal.v7i3.9812engaged within the text: expansion and contraction.In expansion, a text writer actively opens dialogic space with alternative voices; while in contraction such space is challenged and restricted.Each of these subsystems is further divided into subcategories, as can be seen in Figure 1.(Martin & White, 2005, p. 134) As indicated in Figure 1 above, each subsystem of contraction and expansion contains sub-categories with their distinctive lexicogrammatical features.In the contraction system where alternative voices and positions are challenged and restricted, disclaim and proclaim can be used, each of which is further divided into (1) deny and counter, and (2) concur, pronounce and endorse respectively.On the other hand, in the expansive system, which seeks to expand dialog with other voices, there are entertain and attribute.With attribute, further resources to use are acknowledge and distance.As shown in Figure 1 above, each subcategory in both systems of contract and expand uses different lexicogrammatical resources.'Deny', for instance, a subcategory from the contract system, uses negation, while 'entertain' from the expansive system mainly uses modality.
This system of engagement can be used to create an authorial voice that allows 'space' for writers to position themselves, while also opening up alternative voices present in their texts.As Martin and White (2005) maintain, the use of entertain resources, for instance, invokes other voices regarding the proposition being advanced.In this way, the text producer anticipates potential disagreement on the part of the addresee regarding the proposition.This resource can thus be used by writers to create a sense of solidarity with their readers.Another subcategory of expansive resource is 'attribute', which is further divided into two subcategories: acknowledge and distance.In the former, writers merely acknowledge other voices without overtly stating their positions; while in the latter, they explicitly distance "the authorial voice from the attributed material" (p.113).
Even though the use of expansive resources is important to create a dialogic backdrop for a proposition through the inclusion of other voices other than that of the writer's, the resources of dialogic contraction are important in creating strong authorial voice to show the writer's stance on the issue being presented.In 'deny', the first subtype of disclaim for example, writers disalign themselves from other voices presented in their texts and draw their readers to take on their (the writers') positions.The disaligning from alternative voices can also be seen in the second subtype of disclaim, 'counter', which "invokes a contrary position which is then said not to hold" (p.121).
Another group of resources within the contractive system of Engagement that can be used to show an authorial voice is that of proclaim.Unlike the first group ('disclaim'), the resources under 'proclaim' -concur, endorse and pronouncerestrict the dialogic scope of alternative voices instead of directly rejecting them.In 'concur', writers do this by using linguistic resources such of course, naturally, and admittedly to position their readers as having the same view as theirs.Meanwhile, in 'endorse', writers restrict other voices by showing their support to certain voices presented in the text.Finally, in 'pronounce', writers limit the scope of alternative voices through explicit intervention in the textwith phrases such as but the fact of the matters are that or it is absolutely clear to me that.
Adept writers draw on resources from both expansive and contractive systems skillfully in accordance with the type of text they produce.These engagement resources can be a useful tool to analyze voice in both spoken and written texts.In the present study, the aforementioned resources have been drawn on to analyze the construction of authorial voice in students' analytical exposition texts.

Analytical Exposition
In SFL tradition, persuasive writings can manifest in three kinds of genres: (1) analytical exposition, (2) hortatory exposition, and (3) discussionalso known as argumentative text (Emilia & Christe, 2013;Gerot & Wignell, 1994;Humphrey, Droga, & Feez, 2012;Martin, 1989).As traditionally known, exposition presents a one-sided argument about something (Love & Humphrey, 2012); analytical exposition, which is the focus of this study, persuasively argues that something is the case (Gerot & Wignell, 1994;Love & Humphrey, 2012).In this way, analytical exposition differs from hortatory exposition in that the former is only concerned with presenting an argument or point of view without requiring the readers to do something.Analytical exposition is also different from discussion as the latter argues for a case by presenting two or more viewpoints (Emilia & Christe, 2013;Martin, 1989).In short, analytical exposition is a kind of persuasive writing that persuades its audience about the validity of an argument using one-sided perspective.
The social function of analytical exposition is manifested in its generic structure and language features.The structure is relatively straightforward and is common among essay writers: Thesis, Arguments and Reinforcement of Thesis.Gerot and Wignell (1994) state that the Thesis introduces the topic and the writer's position.This part also outlines the main points to be presented, which are restated and developed in the Arguments.Meanwhile, the Reinforcement of Thesis restates the writer's position.In terms of language features, elements that stand out include the focus on generic participants, the use of simple present tense, internal conjunction to stage arguments, and causal conjunction and nominalisation for reasoning.

Engagement and Voice: Previous Studies
Various studies across disciplines have indicated the usefulness of Engagement for analyzing writers' voice.Presently, however, this device has not yet been much drawn on in the analysis of voice in academic writing (see Hyland, 2002Hyland, , 2008;;Ivanic & Camps, 2001;Matsuda, 2001).Within the Appraisal Theory, while Engagement has been used in studies involving legal contexts (Körner, 2000), and police interrogation (Chuanyou, 2008), this notion is rarely the focus of research in academic writing (Coffin, 1997).Among the few studies in academic writing that focused on Engagement involved the analyses of research articles (Fryer, 2013;Hood, 2010;Yuchen, Xuan, & Rui, 2014) and argumentative essays (Liu, 2013;Mei, 2007;Mei & Allison, 2003, 2005), leaving other text types such as analytical exposition almost under-researched.Needless to say, analytical exposition is a very important genre to acquire and develop, especially for language learners, as it provides them with a basis for developing more complex writing skills (Rothery, 1985).In line with this argument, Martin (1989) also believes that competence in writing analytical exposition is highly regarded at the university level.
One of the few research studies investigating writers' voice through Engagement resources in academic texts comes from Prasetyo (2011), who compared the Appraisal of Ahmadiyah issue in the editorial and opinion columns in The Jakarta Post and The Jakarta Globe newspapers.The data involved three text types: hortatory expositions (editorial in The Jakarta Globe), analytical expositions (editiorial in The Jakarta Post), and discussion (opinion texts of both papers).Although not comprehensive, the study showed that in all the text types monogloss was predominantly used by the writers in the two newspapers.The dominant use of this Engagement resource was used by the authors to show their authorial voice on the issue at hand, which was to side with the Ahmadiyah followers.
Another study that featured analytical exposition was conducted by Pusparini (2014), who analyzed the Appraisal of four opinion texts on the banning of polygamy in The Jakarta Post.Similar to Prasetyo's study above, analytical exposition in this study was just a text type that was used as the data, in addition to hortatory exposition and discussion.The findings revealed that the dominant Engagement resource used in all the texts was, again, monoglossic.The dominant use of this particular Engagement resource seemed to have successfully created strong authorial subjective voice towards the issue of polygamy banning in the texts involved.
Yet, other studies have indicated that the use of Engagement resources in writing is highly influenced by the type of text being produced.Analyzing Engagement in argumentative essays on geography, for instance, Mei (2007) showed how high-rated essays in the study combined both monoglossic and hetereglossic resources of Engagement to create strong authorial voice suitable for the text type.Meanwhile, Liu (2013), who also analyzed voice and Engagement in argumentative writings, found that less frequent use of monoglossic resources in low-graded essays created weaker authorial voice.In addition, low graded essays in her study were reported to use more contractive resources of disclaim, which made the voice less persuasive.In another study investigating Engagement in medical research articles, Fryer (2013), pointed out that the dominant resource found in his data was 'entertain', which indicates the typical voice of medical science writing as being "cautious, modest or lacking assertions" (p.198).His study also revealed that hetereglossic resources were found more in Introduction and Discussion while monoglossic in Method and Results, indicating a more dialogic nature to alternative voices in the first two sections.
Considering the little attention given to analytical exposition in studies of Engagement and voice, especially in EFL (English as a Foreign Language) settings, this study examines how Indonesian university students of English study program constructed writers' voice in analytical exposition texts.Further details of the participants and the context are given in the Method section below.

METHOD
Seeking to analyze how writers' voice is constructed through Engagement resources, this study collected nine analytical exposition texts written by fourth semester university students for a literary course.While twenty-three students were enrolled for the course, only nine of the students' texts were selected.The choice has been purposeful, taking the students' English proficiency levels and writing competence as the main criteria for selection.Prior to the selection, the twenty-three students' writings had been categorized into three: above average (AA), average (A), and below average (BA).This classification was particularly helpful in that it allowed better insights to be sought into how each group established their writers' voice through their selection of Engagement resources.Due to the limited scope of the study, however, only three texts from each student group were used as data.Overall, these texts revealed the students' views regarding the importance of literature, which was based on a previously given reading material entitled 'Why literature matters' written by Gillespie (1994).Martin and White's Engagement system (Martin & White, 2005) was employed as both theoretical and methodological devices for analyzing writers' voice conveyed in the texts.This framework was used because, as the literature review above suggests, the framework allows a rigorous analysis of Engagement resources in texts.It is worth noting that, in the present study, It is worth noting that, in the present study, sentences are the main unit of analysis, and --since grammatical accuracy is not the concern of the study--grammatical errors are not discussed.

Voice and Engagement in Students' Texts
Data analysis revealed some similarities and differences in terms of Engagement patterns in the students' texts, which in turn influence the construction of voice in their writings.However, due to space restriction, only major trends will be discussed in this section.
In terms of similarity, all of the students' texts were heteroglossically expansive.The heteroglossic nature of these texts were indicated in other voices that the writers drew on to present their case.Through expansive resources, these writers showed that they acknowledged alternative voices and thus expanded the dialogic space in their texts.The heteroglossic nature of the texts in this study is not surprising since the prompt of the task itself features an external view from Gillespie (1994) regarding 'why literature matters'; therefore, from the outset, a heteroglossic backdrop for the development of the argument had been created.The use of heteregloss in the students' texts is in line with Bakhtinian's notion of dialogism (see Martin & White, 2005), which maintains that all texts must be related to and in some way 'echo' other texts.
Data analysis also revealed that the dominant dialogic expansion resources used in this study are 'entertain' and 'acknowledge'.The students' preference for incorporating 'entertain' resources in their texts suggests their recognition for alternative voices on the issue at hand.As argued by Martin and White (2005), 'entertain' also serves to create solidarity with their putative readers who might have different opinions from them.The dominant use of expansive resources, especially 'entertain', in the students' texts in this study is similar to the Engagement patterns of medical research articles, which have been characterized as being "cautious, modest, or lacking assertion" (Fryer, 2013, p. 198).
Meanwhile, 'acknowledge' was used by the students to source external voices without overtly aligning or disaligning their stance with these voices.This strategy seemed to have been adopted by the writers to provide a preliminary background to their argument.Students from the three groups used this resource in collaboration with other Engagement resources to present their voicewith differing levels of successas exemplified in Excerpt 1 below.As can be seen in Excerpt 1 above, Student A3 used 'acknowledge' along with other resources of Engagement.The underlined part of the excerpt is a resource of acknowledge that takes form in the noun phrase the main argument.Through this engagement resource, the writer reports Gillespie's main arguments without overtly taking side with them.The writer used this external-source proposition as an introduction to the argument that he was yet to present in the rest of the paragraph.Using other Engagement resources, namely 'pronounce', 'endorse', and 'entertain', this student successfully presented his case in an affirmative tone.Through the use of the 'pronounce' "and I agree with this statement", Student A3 explicitly shows his stance on the previously cited external proposition.To align the readers to his stance, both contractive and expansive resources of Engagement were used.The contractive resource used in this paragraph is 'endorse' ("and it also teaches me about living a life in the world in many ways").The contractive resource was employed to strengthen the writer's authorial voice on his support to Gillespie's proposition which is realized through the word "teaches" that endorses the idea about the importance of literature.Meanwhile, the expansive Engagement resources used in this paragraph is 'entertain', i.e., "I can imagine that I become someone else and living a life with a different problem or maybe same, I will have a hard-life starter pack, I will appreciate more my real life, I will know which is good and which is bad, and I will get a lot of experiences to have a different perspective to live my life".Each of these resources indicates one possible way of seeing the matter regarding the proposition being advanced in the first two clauses of the paragraph, thus entertaining other possibilities in seeing the matter.The dominant expansive resources of Engagement used in this paragraph created a sense of dialogue with the readers, which makes the proposition being advanced non-threatening, while the use of contractive resources created a clear authorial stance.The use of these two Engagement resources in the paragraph resulted in the establishment of authorial voice which is proper for an analytical exposition text.
However, the study also indicated that the use of Engagement resources did not always bear similar successful results.While average and above average students consistently used 'acknowledge' in collaborative harmony with other resources to create an authorial voice, below average student did not.This is exemplified in Excerpt 2 below taken from Student BA2.The engagement resources employed by BA2 in the excerpt above did not create a strong authorial voice due to lack of elaboration in each Engagement resource used.Dominated by 'acknowledge', the writer presented his propositions from external sources without showing his side towards them (see Martin & White, 2005), e.g., through the use of reported verb said.In addition, the lack of elaboration of the propositions being advanced and the absence of clear logical connection between one proposition and the other further blurs the clarity of the authorial voice in this paragraph.Thus, even though another Engagement resource of 'concur' is used at the end of the paragraph to present his voice-which was also used by medical research article writers in Fryer's study (2013) to "affirm the validity of the claims" (p.200)-it did not create such a succesful result in this particular paragraph.As shown in the excerpt above, 'concur' was used by BA2 by first invoking dialog through the question "Who needs literature?," and then restricting the dialog by providing the answer ("We all do.") to the question posed.However, due to lack of elaboration in the previous sentences, such effort was not too successful in establishing a strong authorial voice.
Another point to highlight is that none of the texts in the data used 'distance'.As Martin and White (2005) argue, writers use this resource to dissociate themselves from the stance they include in their texts.The absence of this resource in the data is not surprising due to the nature of analytical exposition texts, which convey a one-sided view of a case without having to present a counter argument to it (Emilia & Christe, 2013;Gerot & Wignell, 1994;Humphrey et al., 2012;Love & Humphrey, 2012;Martin, 1989).Regarding this, Fryer (2013) also argues that writers of milder persusive texts such as research articles, which was the focus of his study, might find this Engagement resource to be too strong.'Distance' resources are commonly adopted in writing a persuasive genre known as hortatory exposition, where differing viewpoints are challenged and countered, as can be found in texts like Letters to the Editor (Fryer, 2013).
Apart from the above-mentioned similarities, the study also revealed differences in the use of Engagement strategies by the three groups of students.While all of the students' texts were heteroglossic in nature, none of the below average (BA) students used monogloss in their texts.A study conducted by Liu (2013) also showed a similar tendency, in which high-rated essays in her study used more monoglossic resources than the low-rated ones.Since monogloss in a text is important to create a strong authorial stance (Martin & White, 2005), making it more "affirmative and authoritative" (Liu, 2013, p. 47), the absence of this Engagement strategy thus reduces the sense of authorship on the part of the writer.To compensate for the absence of monogloss, below average (BA) students in this study relied on contractive resources instead: 'deny' and 'counter' (used by BA1), 'affirm' (BA2), 'counter' (BA3) and 'pronounce' (used by all three).As argued by Mei and Allison (2005), contractive resources of Engagement serve to create finality to dialogic nature of texts.The following excerpt taken from the writing of BA2 in Excerpt 2 above illustrates this point: Excerpt 3 Who needs literature?We all do (BA2 -'concur', 'affirm') This student used the Engagement strategy of 'concur', in this case 'affirm', to present his case regarding the importance of literature.Related to this strategy, Martin and White (2005) point out that 'concur' allows the writers to align the reader with their position.In this particular example, the student did this by first posing a rhetorical question, which served to invoke many voices, and then restricted the previously invoked alternative voices by affirmatively stating "we all do".By using the pronoun we, the writer positioned his readers to be in alignment with his view, hence restricting the dialogic space for alternative voices.

CONCLUSION
This study has shown that students from the three groups of proficiency levels are capable of exploiting the resources of Engagement to create an authorial voice in their analytical exposition texts, despite with differing levels of success.The dominant heteroglossically expansive Engagement pattern of their texts indicates these students' acknowledgement of alternative positions.
It has also been shown that even though all the texts were generally heteroglossic, the resources of Engagement used by students from each group of proficiency level was different.Students from the below average (BA) group mainly used expansive resources to entertain possible voices on the issue at hand, which resulted in a weaker authorial voice.In addition to relatively varied types of expansive resources, average (A) group students also used contractive resources of endorsement to support their voice.However, the dominant use of 'attribution' to merely acknowledge other voices made the writers' voice from this group loosely connect to external voices.Meanwhile, above average (AA) students' texts are characterized by the writers' well exploitation of both types of heteroglossic resources of Engagement, in particular that of 'proclaim' from contractive resources, which were used to explicitly show their position on the issue at hand.This resulted in a stronger writers' voice, supported by the external voices they included in their texts.
The study has also confirmed the findings of previous studies (e.g., Liu, 2013) that texts with a strong authorial voice are not those that use exclusively heteroglossic Engagement patterns, but rather those that show an interplay between the two patterns of Engagement (monogloss and heterogloss), including that within the complex subsystems of heterogloss.Future researcher interested in conducting investigation in the same area may consider analyzing analytical exposition texts from different subject areas outside literature to see the pattern of Engagement resources used for voice construction.Future research may also involve data in the form of spoken texts.The present study suggests pedagogical implications for the teaching of writing: students need to be taught the importance of voice and stance in their writing in order to construct a text with a strong sense of authorship.As the analysis of the students' writing in this study indicates, successful writers are able to skillfully draw on a range of Engagement resources and strategies in their writing.It follows that 'voice', along with its other relevant notions such as dialogism and heteroglossia, needs to be integrated into the Writing syllabus as part of theoretical content knowledge within the course.Additionally, as the present study suggests, it is necessary that features and functions of Engagement resources (see Figure .1) be explicitly taught in the classroom.In teaching analytical expositions, for instance, students can be taught how lexicogrammatical resources in the contractive and expansive Engagement subsystems can be exploited to tone down or strengthen their own voices.Similarly, students can be taught how to employ these resources to align or disalign their voices with those of others.In short, it is of high importance that teachers be able to explicitly show their students how to draw on Engagement resources and strategies appropriately in order to construct a wellargued text.On top of this, ample opportunity has to be provided for students to practice and develop their writing skills and, more importantly, that constructive feedback be given by the teacher.It is of great hope that this study can lead to a betterinformed instructional practice.